3D Shapes Local Geometry Codes Learning with SDF

被引:11
|
作者
Yao, Shun [1 ,2 ]
Yang, Fei [1 ,2 ]
Cheng, Yongmei [1 ,2 ]
Mozerov, Mikhail G. [2 ]
机构
[1] Northwestern Polytech Univ, Sch Automat, Xian, Peoples R China
[2] Univ Autonoma Barcelona, Comp Vis Ctr, Barcelona, Spain
关键词
D O I
10.1109/ICCVW54120.2021.00239
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
A signed distance function (SDF) as the 3D shape description is one of the most effective approaches to represent 3D geometry for rendering and reconstruction. Our work is inspired by the state-of-the-art method DeepSDF [17] that learns and analyzes the 3D shape as the iso-surface of its shell and this method has shown promising results especially in the 3D shape reconstruction and compression domain. In this paper, we consider the degeneration problem of reconstruction coming from the capacity decrease of the DeepSDF model, which approximates the SDF with a neural network and a single latent code. We propose Local Geometry Code Learning (LGCL), a model that improves the original DeepSDF results by learning from a local shape geometry of the full 3D shape. We add an extra graph neural network to split the single transmittable latent code into a set of local latent codes distributed on the 3D shape. Mentioned latent codes are used to approximate the SDF in their local regions, which will alleviate the complexity of the approximation compared to the original DeepSDF. Furthermore, we introduce a new geometric loss function to facilitate the training of these local latent codes. Note that other local shape adjusting methods use the 3D voxel representation, which in turn is a problem highly difficult to solve or even is insolvable. In contrast, our architecture is based on graph processing implicitly and performs the learning regression process directly in the latent code space, thus make the proposed architecture more flexible and also simple for realization. Our experiments on 3D shape reconstruction demonstrate that our LGCL method can keep more details with a significantly smaller size of the SDF decoder and outperforms considerably the original DeepSDF method under the most important quantitative metrics.
引用
收藏
页码:2110 / 2117
页数:8
相关论文
共 50 条
  • [1] Learning Material-Aware Local Descriptors for 3D Shapes
    Lin, Hubert
    Averkiou, Melinos
    Kalogerakis, Evangelos
    Kovacs, Balazs
    Ranade, Siddhant
    Kim, Vladimir G.
    Chaudhuri, Siddhartha
    Bala, Kavita
    2018 INTERNATIONAL CONFERENCE ON 3D VISION (3DV), 2018, : 150 - 159
  • [2] Geometry curves: A compact representation for 3D shapes
    Li, Guo
    Liu, Ligang
    GRAPHICAL MODELS, 2013, 75 : 265 - 278
  • [3] VRMath: A 3D microworld for learning 3D geometry
    Yeh, A
    Nason, R
    ED-MEDIA 2004: WORLD CONFERENCE ON EDUCATIONAL MULTIMEDIA, HYPERMEDIA & TELECOMMUNICATIONS, VOLS. 1-7, 2004, : 2183 - 2191
  • [4] Geometry image matching for similarity estimation of 3D shapes
    Laga, H
    Takahashi, H
    Nakajima, M
    COMPUTER GRAPHICS INTERNATIONAL, PROCEEDINGS, 2004, : 490 - 496
  • [5] Geometry and Convergence Analysis of Algorithms for Registration of 3D Shapes
    Helmut Pottmann
    Qi-Xing Huang
    Yong-Liang Yang
    Shi-Min Hu
    International Journal of Computer Vision, 2006, 67 : 277 - 296
  • [6] Geometry and convergence analysis of algorithms for registration of 3D shapes
    Pottmann, Helmut
    Huang, Qi-Xing
    Yang, Yong-Liang
    Hu, Shi-Min
    INTERNATIONAL JOURNAL OF COMPUTER VISION, 2006, 67 (03) : 277 - 296
  • [7] Watermarking 3D shapes using local moments
    Bors, AG
    ICIP: 2004 INTERNATIONAL CONFERENCE ON IMAGE PROCESSING, VOLS 1- 5, 2004, : 729 - 732
  • [8] 3D GEOMETRY AND LEARNING OF MATHEMATICAL REASONING
    Mithalal, Joris
    CERME 6 - PROCEEDINGS OF THE 6TH CONGRESS OF THE EUROPEAN SOCIETY FOR RESEARCH IN MATHEMATICS EDUCATION, 2010, : 796 - 805
  • [9] Learning Line Features in 3D Geometry
    Sunkel, M.
    Jansen, S.
    Wand, M.
    Eisemann, E.
    Seidel, H. -P.
    COMPUTER GRAPHICS FORUM, 2011, 30 (02) : 267 - 276
  • [10] Learning by Restoring Broken 3D Geometry
    Liu, Jinxian
    Ni, Bingbing
    Chen, Ye
    Yu, Zhenbo
    Wang, Hang
    IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 2023, 45 (09) : 11024 - 11039